initContainer fails on CRD upgrade job
Bug
Description
When running the CRD upgrade job, the init container is supposed to move the kubectl binary into /tmp to mount it and make it available to the velero image that actually pushes out the CRDs into the cluster. The problem is it assumes the kubectl binary is in /usr/local/bin/kubectl
. However, the image it is trying to do that on (registry1.dso.mil/ironbank/opensource/kubernetes/kubectl:1.27.6) actually puts that in /opt/bitnami/kubectl/bin/kubectl
. So the initContainer persistently fails since the command can't copy a non-existent file into /tmp.
BigBang Version
I'm just using the chart, but pulling specifically the tag 5.2.2-bb.1
.